Affine merge mode with translational motion vectors

Abstract

In some examples, an apparatus for video decoding includes processing circuitry. The processing circuitry determines, from a candidate list, a first affine-translational merge candidate for a prediction of a current block in a current picture. The first affine-translational merge candidate provides affine motion information associated with a first reference picture in a first reference list and translational motion information associated with a second reference picture in a second reference list. The processing circuitry generates a first prediction for a sample in the current block according to the affine motion information associated with the first reference picture, and generates a second prediction for the sample in the current block according to the translational motion information associated with the second reference picture. The processing circuitry reconstructs the sample of the current block according to a combination of the first prediction and the second prediction.

What is claimed is: 
     
         1 . A method of video processing in a decoder, comprising:
 receiving a coded video bitstream comprising a current block in a current picture;   determining, from a candidate list, a first affine-translational merge candidate for a prediction of the current block in the current picture when a high level syntax in the coded video bitstream is determined to indicate that the first affine-translational merge candidate for the prediction of the current block is allowed, the first affine-translational merge candidate providing affine motion information associated with a first reference picture in a first reference list and translational motion information associated with a second reference picture in a second reference list, the first affine-translational merge candidate providing different types of motion information associated with different reference pictures, the first reference picture being different from the second reference picture, the translational motion information associated with the second reference picture being non-symmetric to the affine motion information associated with the first reference picture;   generating a first prediction for a sample in the current block according to the affine motion information associated with the first reference picture;   generating a second prediction for the sample in the current block according to the translational motion information associated with the second reference picture; and   reconstructing the sample of the current block according to a combination of the first prediction and the second prediction.   
     
     
         2 . The method of  claim 1 , further comprising:
 deriving the affine motion information and the translational motion information of the first affine-translational merge candidate; and   inserting the first affine-translational merge candidate in the candidate list.   
     
     
         3 . The method of  claim 2 , wherein the deriving the affine motion information comprises at least one of:
 deriving the affine motion information according to first affine motion information of a first affine merge candidate with uni-prediction; and   deriving the affine motion information according to second affine motion information of a second affine merge candidate with bi-prediction, the second affine motion information being associated with the first reference list.   
     
     
         4 . The method of  claim 2 , wherein the deriving the translational motion information comprises at least one of:
 deriving the translational motion information according to first translational motion information of a first translational merge candidate with uni-prediction; and   deriving the translational motion information according to second translational motion information of a second translational merge candidate with bi-prediction, the second translational motion information being associated with the second reference list.   
     
     
         5 . The method of  claim 2 , wherein the deriving the translational motion information comprises:
 for a merge candidate with a merge index in a translational merge candidate list,
 selecting a first list from the first reference list and the second reference list based on the merge index according to a predefined rule; 
 deriving the translational motion information based on first translational motion information of the merge candidate associated with the first list when the first translational motion information exists; and 
 deriving the translational motion information based on second translational motion information of the merge candidate associated with another list different from the first list when the first translational motion information does not exist. 
   
     
     
         6 . The method of  claim 2 , wherein the deriving the translational motion information comprises at least one of:
 deriving the translational motion information based on a control point motion vector of an affine merge candidate;   deriving the translational motion information based on an average of a plurality of control point motion vectors of the affine merge candidate; and   deriving the translational motion information based on a motion vector at a center of the current block, the motion vector being determined according to an affine model from the affine merge candidate.   
     
     
         7 . The method of  claim 2 , wherein the inserting the first affine-translational merge candidate in the candidate list further comprises at least one of:
 inserting the first affine-translational merge candidate in an existing subblock based merge candidate list after a last inherited affine merge candidate;   inserting the first affine-translational merge candidate in the existing subblock based merge candidate list after a last constructed affine merge candidate; and   inserting the first affine-translational merge candidate in the existing subblock based merge candidate list after a last affine merge candidate and before a zero motion vector candidate.   
     
     
         8 . The method of  claim 2 , further comprising:
 inserting the first affine-translational merge candidate into an affine-translational merge list separate from a subblock based merge candidate list.   
     
     
         9 . The method of  claim 8 , wherein the determining the first affine-translational merge candidate comprises:
 when a decoded flag from the coded video bitstream indicates a selection of the affine-translational merge list, selecting the first affine-translational merge candidate from the affine-translational merge list according to a decoded index from the coded video bitstream.   
     
     
         10 . The method of  claim 8 , further comprising:
 reordering candidates in the affine-translational merge list according to template matching costs of the candidates.   
     
     
         11 . The method of  claim 2 , wherein the inserting the first affine-translational merge candidate in the candidate list further comprises:
 inserting the first affine-translational merge candidate in the candidate list when a counting number of affine-translational merge candidates in the candidate list is lower than a predefined value.   
     
     
         12 . The method of  claim 2 , wherein the deriving the affine motion information comprises:
 deriving the affine motion information from a subset of affine merge candidates that is in front of other affine merge candidates in an affine merge candidate list, a counting number of affine merge candidates in the subset being equal to a predefined value.   
     
     
         13 . The method of  claim 2 , wherein the deriving the translational motion information comprises:
 deriving the translational motion information from a subset of translational merge candidates that is in front of other translational merge candidates in a translational merge candidate list, a counting number of the translational merge candidates in the subset being equal to a predefined value.   
     
     
         14 . The method of  claim 1 , further comprising:
 making a determination that bi-prediction is allowed in the current picture; and   determining the first affine-translational merge candidate for the prediction of the current block when the determination is made.   
     
     
         15 . The method of  claim 1 , further comprising:
 making a determination that reference pictures of the current picture are temporally before the current picture; and   determining the first affine-translational merge candidate for the prediction of the current block when the determination is made.   
     
     
         16 . The method of  claim 1 , further comprising:
 when local illumination compensation (LIC) is enabled, setting a LIC flag associated with the first affine-translational merge candidate to false.   
     
     
         17 . The method of  claim 1 , wherein the reconstructing the sample of the current block further comprises:
 setting a bi-prediction with coding unit level weighting (BCW) index to a default value for combining the first prediction and the second prediction with equal weighting.   
     
     
         18 . The method of  claim 1 , wherein the reconstructing the sample of the current block further comprises:
 determining that one of the affine motion information and the translational motion information has a bi-prediction with coding unit level weighting (BCW) index; and   combining the first prediction and the second prediction according to the BCW index.   
     
     
         19 . The method of  claim 1 , wherein the reconstructing the sample of the current block further comprises:
 determining that a first bi-prediction with coding unit level weighting (BCW) index is associated with the affine motion information and a second BCW index is associated with the translational motion information; and   combining the first prediction and the second prediction according to the first BCW index.